INDEX

Explanations

references to President Trump and other political figures

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 21-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L21R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.21.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Wich

-0.17

 Kral

-0.15

ired

-0.14

ANGES

-0.14

itech

-0.14

 minister

-0.14

sup

-0.14

RTL

-0.13

ite

-0.13

 wake

-0.13

POSITIVE LOGITS

 Barack

0.26

 Obama

0.24

 Trump

0.22

 Bush

0.20

 President

0.19

 Donald

0.19

Obama

0.18

 Clinton

0.17

 Biden

0.17

Trump

0.17

Activations Density 0.068%

references to President Trump and other political figures

No Comments

No Known Activations